2013

Jennifer Widom
Summer 2013

Dear Computer Science Alumni and Friends,

This past year hasn’t been quite as tumultuous as the previous one, when consideration of a New York City campus and rapid-fire developments in online education kept all of us on our toes (and occasionally awake at night). On the other hand, it’s not as if we’ve been sitting around with nothing to do. CS106A became so popular the fire marshal showed up to shoo students from the aisles, the number of CS majors now dwarfs every other major on campus, and our growing computer clusters in the basement of the Gates building regularly blow out the electricity. We’re bursting at the seams, but also reveling in our success.

The past several years have been especially fruitful ones in terms of faculty hiring. Last year alone garnered four new professors covering a spectrum of computer science: computational biology, data management, computer vision, and theory. Our new faculty members are introduced below. This newsletter also describes several successful new programs supporting women in CS from high school to PhD, and describes a variety of projects where CS students have put their talents to work in aid of society. As usual, our faculty, staff, and students have gathered a large number of awards and honors. A few research synopses and an update on online education round out the newsletter.

Please enjoy the many departmental highlights of the 2012-13 academic year.

New Faculty and Lecturers

We’re very excited to have four new faculty joining the department, which might just be a record for a single year. They cover a wide swath of research areas and, despite their "junior" status, already bring with them an unusual wealth of research experience.

• Ron Dror received his Ph.D. from MIT in 2002, and since then has been working at D.E. Shaw Research. “DESRES” was founded by Stanford CS Ph.D. alum David Shaw, but should not be confused with the well-known D.E. Shaw investment firm—DESRES is a thriving research lab in the area of computational biochemistry. Ron was DESRES’ first hire, and he has been serving in a leadership role throughout his time at the lab. Ron is excited to move to the academic environment and ramp up his own research program in computation and biology. We’re particularly excited that Ron is jointly appointed between Computer Science and the Institute for Computational and Mathematical Engineering (ICME). Although CS and ICME have a long history of collaborations, Ron will further strengthen those important ties.
• Christopher Ré received his Ph.D. from the University of Washington in 2009, and he joins us after four years on the faculty at the University of Wisconsin. Chris sums up his already voluminous body of work in databases, machine learning, and “Big Data” in general—along with his plans for the future—with a simple statement on his web page: I believe that the future of computing is in data management. (As a database researcher myself, I’m not going to argue!) In addition to his fundamental theoretical and practical work, Chris has fostered a number of interesting interdisciplinary collaborations, including a project whose code is deployed at the South Pole as part of the IceCube Neutrino Detector. We can’t wait to see what Chris does at Stanford.
• Silvio Savarese received his Ph.D. from CalTech in 2005, spent three years as a Beckman Institute Fellow at the University of Illinois, then joined the EECS faculty at the University of Michigan before moving to Stanford this fall. Silvio brings strength and experience in the very important area of computer vision. Silvio’s work specifically focuses on analyzing and modeling visual scenes from static images and video sequences, with the goal of enabling machines to perform a variety of real-world visual tasks. In his hiring talk, Silvio showed a short video of his very young son crawling around their home, effortlessly grabbing and manipulating a variety of objects. His research objective is to help computers become as adept as young Leo.
• Virginia Vassilevska Williams  received her Ph.D. from Carnegie Mellon in 2008. Since then she has been a post-doc at the Institute for Advanced Study (Princeton) and UC Berkeley, and a research scientist jointly appointed between UC Berkeley and Stanford. Virginia is a theorist—the latest in a recent spate of hiring that has quickly catapulted our theory group from a modest few (but excellent) faculty, to one of the strongest groups anywhere. Virginia is particularly well-known for a dramatic result showing the first improvement in 25 years to the known complexity upper-bound for matrix multiplication. It may not appear dramatic that the improvement comes in the form of an exponent reduced from 2.376 to 2.373, but if you think about how often matrices are multiplied, even a tiny improvement quickly adds up! In her research scientist position Virginia has already been conducting exciting research with Stanford students and professors; we’re thrilled to have her join the regular faculty.

In addition to four new faculty, to meet the rapidly increasing demand for CS courses (more on this topic shortly), we’re adding two new Lecturers to our excellent staff:

• Cynthia Lee joins us from UC San Diego, where she received her CS Ph.D. in 2009, and has been a Lecturer in the Computer Science and Engineering department since 2007. Joining Stanford is a homecoming for Cynthia, who grew up in the Bay Area.
• Marty Stepp joins us from the University of Washington, where he has been a Lecturer and Senior Lecturer in the Computer Science and Engineering Department since 2006. Marty is an author of three CS textbooks, including one of the leading introductory Java textbooks currently available.

We're very excited to have Cynthia and Marty adding further strength to our world-renowned Computer Science Education program.

Quiz follow-up: Last year's newsletter mentioned that new faculty member Michael Bernstein became our third professor to have participated in the much-loved CS section leader program as an undergraduate, and asked if anyone knew who the other two are. Responses were, shall we say, slim, but for completeness the answer is:  Tim Roughgarden and Mehran Sahami.

Tenure and Retirement

I'm pleased to announce that Prof. Phil Levis and Prof. Fei-Fei Li have been promoted to Associate Professor with tenure. We're very excited to enjoy Phil and Fei-Fei's considerable contributions to the research and teaching missions of the department for many years to come.

Meanwhile, Prof. Terry Winograd has retired his regular faculty position, although as with many Emeritus faculty, Terry continues to stay closely involved with the department. Throughout his career, Terry's research and teaching have been truly at the vanguard of major advances: from artificial intelligence to social computing, ubiquitous computing, and design thinking. Terry’s Ph.D. thesis was a landmark in natural language processing. Terry went on to make seminal contributions to human-computer interaction (HCI) and social computing through his books Understanding Computers and Cognition and Bringing Design to Software. In 1990, Terry founded Stanford’s program on People, Computers, and Design—one of the world’s first and most visible academic HCI programs. The program’s Web broadcast seminar has run nonstop for 23 years. Terry’s Interactive Workspaces project was an early and ambitious effort exploring computing beyond the desktop. Terry has educated an entire generation of computer scientists, including many successful entrepreneurs (e.g., CEO Larry Page) and academics (e.g., Cornell's Francois Guimbretiere); you can watch their toasts to Terry at the HCI group’s 20th anniversary celebration. Recently, Terry was one of the founding faculty of the Hasso Plattner Institute of Design at Stanford (better known as the “d.school”), teaching design thinking to students campus-wide. Last but not least, with his curly gray locks and shining intellect, Terry bears a striking resemblance to Albert Einstein, yielding delighted photograph requests from new freshmen to proud parents at graduation.

CS Popularity Shows Sharp Increases for Fifth Year in a Row

This is the fifth year in a row with significant increases in both CS major declarations and CS course enrollments. Although I’m beginning to sound like a “broken record” (hmm, does the iPod generation even use that phrase?), we set another all-time high for the number of undergraduates choosing to major in CS: 261 students declared CS as their major this past year alone. Computer Science remains the most popular undergraduate major at Stanford, now a whopping 50% larger than the second-most-popular major, Human Biology.

In addition to record numbers of CS majors, the number of students from across the university choosing to take at least one CS courses is unprecedented. More than 2000 students took one of the entry-level programming courses this past academic year (the vast majority enrolling in CS106A), a number that’s far higher than the size of each year’s freshman class. How can that be, you ask? In addition to nearly all Stanford undergraduates opting to take at least one CS class, quite a few graduate students in other fields are taking advantage of their time at Stanford to learn about CS as well.

Classrooms of several hundred students are becoming commonplace in CS. We spend lots of time sharing techniques for managing and teaching such large classes effectively. It’s a continuing challenge, but at the same time it’s hard to complain about the immense popularity of our department.

Women in Computer Science

While the percentage of women CS majors may seem discouragingly low, we’re actually well above the national average, and we’re especially pleased that our percentage is rising steadily. In 2012, 18% of our BS degrees were awarded to women (against a national average of 13%), and this year we're up to 23%. Of course we still have some distance to go. CS students and faculty have launched a variety of programs to encourage women at all levels to embrace Computer Science, varying from one-day boot camps exposing high school girls to programming for the first time, to research exchanges for PhD students. Some of our programs have gained national visibility. Here’s a sampler:

• she++. When two of our undergraduates came up with the name “she++” while brainstorming a possible day-long event around women in computing, the name itself was so clever the event had to go forward. It was a huge success, and she++ has transformed from an event into a movement. Described as “a Stanford-based community for innovative women in technology” with the goal of “inspiring women to empower computer science,” a widely-acclaimed documentary video has been produced and more events are planned. The video was released publicly this summer. Find the video and lots more information on the she++ website.
• Girls Teaching Girls to Code. Empirical studies show clearly that encouraging women students to embrace computer science (or any STEM field for that matter) is most effective if it occurs well before college. Several of our undergraduates developed a program where, in one fun-filled tuition-free Saturday, Stanford CS students teach Bay Area high school girls to code. The 2013 event quickly reached capacity of 120 high school girls and 25 mentors, with an equal number on the waiting list. The organizers have another big event planned for 2014, with “mini-events” to keep the momentum going in the meanwhile. For more information and to join the mailing list, visit the Girls Teaching Girls website.
• Girls Gone Geek. For those Bay Area high school girls ready to make a bigger time investment in computer science, thanks to an anonymous donor we launched our inaugural "Girls Gone Geek" two-week summer programming workshop. Organized by Prof. Steve Cooper, the workshop is underway at the time of this newsletter writing, and appears on track to be a great success. More information here.
• Berkeley-Stanford Research Meetup. Finally, for women who have fully embraced computer science and are getting involved in research, CS and EE students at Stanford and Berkeley organized a joint research “meetup” to share their experiences and their research in a supportive and fun environment. The 2013 event was held across the bay, so next year Stanford is on tap to host. Photos and other information from the event can be found on the Research Meetup website.

CS for Social Good

While many of our students are on a direct trajectory towards engineering positions at software companies¾and frankly the market has never been hotter for these students¾there’s a healthy cohort taking a different path: students putting their technical skills to work in direct aid to society. Here are just two examples, on opposite ends of the globe:

• As their classmates danced away at Stanford’s signature all-night dance marathon to raise money in support of international health, CS students huddled at an adjacent all-night “code jam,” producing software for social causes. One team of coders worked on a system for modernizing and streamlining administration of the foster care process; another team worked on a platform to help grassroots organizations in Africa create websites and collaborate with one another; several other projects were underway. The dance-marathon code jams are organized by Code the Change, an organization started by Stanford students whose mission is “to help computer scientists use their skills for social change, and to make social change an integral part of the computer science culture.” Our friends across the bay at Cal have followed our lead, recently starting their own Code the Change chapter.
• Two CS majors recently traveled with a team of four Stanford students to Ethiopia, with the goal of improving conditions in the large refugee camps that turn into de facto cities overnight. Our CS students worked directly in the camps, designing registration systems and communication networks to vastly improve day-to-day life. You can read more about the students and their experiences in a fascinating article that appeared in the Stanford Report: Stanford CS students work with UN on innovation technology for refugees.

The Latest in Online Education

There's no hiding the fact that online-education initiatives have taken the world by storm, with near-daily articles in major newspapers, and just about every institution of higher learning taking careful stock of the opportunities, challenges, and, to be honest, threats, afforded by the extremely rapid developments. And there's no denying that the genesis of the storm was right here in the Stanford Computer Science Department.

As described in last year's newsletter, things actually got started several years ago, with CS department experiments in both course-platform software and "flipped-classroom" teaching. Next came the fateful fall of 2011, when Stanford CS delivered the first set of what's since become known as MOOCs (Massive Open Online Courses): full-length, full-featured courses offered to anyone with an internet connection, for free. Hundreds of thousands of students enrolled in those three inaugural MOOCs, with tens of thousands completing them.  Two startups were launched from our department (and one more from Stanford's Management Science and Engineering department), many more MOOCs have been deployed, and Stanford named CS faculty member John Mitchell as Vice Provost for Online Learning, to oversee initiatives both within campus and beyond. One of those initiatives is to support data-driven research in online-education under the auspices of the new Lytics Lab. Some interesting advances in this area have already been made by CS faculty and students; see the Codewebs project under Research Activities below.

It’s exciting times in online education, with the future still anyone’s guess. To keep abreast of Stanford’s many online-education activities, in the Computer Science department and university-wide, please visit the Stanford Online website.

Research Activities

As always, a tremendous amount of exciting research is going on in the department. Please enjoy these four samples, representing a diverse cross-section of topics and research styles, and visit the CS Department website to learn about our ongoing research across the entire field.

CodeWebs: Providing Complex and Informative Feedback in Programming MOOCs

Prof. Leonidas Guibas, together with postdoctoral researcher Jonathan Huang and graduate students Chris Piech and Andy Nguyen, have been studying in their CodeWebs project how to provide individualized feedback to tens of thousands of students taking massive programming courses on-line. The work is predicated on two simple ideas: (1) When the solution space to a programming assignment is densely sampled by the massive student population, assignments can be meaningfully linked into a network connecting highly related solutions. Such a network provides paths for transporting annotations, comments and corrections, from one solution to another. (2) The majority of erroneous solutions fall into a small number of clusters that are made evident by the network. Human graders can then evaluate one or a few assignments from each cluster, and their comments can be diffused by the network to provide specific feedback to a large number of student solutions. Both the program (tree) matching algorithms involved and the evaluation of consistency of transport along different network paths generate interesting research problems. Unlike crowd-sourced solutions, this approach makes more parsimonious use of human resources and may lead to more fair and consistent results.

Look Who's in the Pool!

Regulars of the Stanford swimming pool will soon be surprised to discover a new denizen: Sinbad, an underwater robot in the form of a human diver. Unlike conventional underwater vehicles, Sinbad has been designed specifically to safely explore fragile coral reef environments. Sinbad's compliant lightweight arms, agile body, stereo vision, and tactile fingers will allow the robot to inspect, collect samples, build structures, and perform manipulation tasks under the guidance of a marine biologist. The robot effectively extends the human’s hands to reach and touch remote fragile environments, achieved through an intuitive user interface, a fast optical communication link with a bimanual haptic device, and an advanced whole-body robot controller.  The haptic device allows the biologist to control the interaction of the robot hands with the marine environment, while the complex robot movement coordination is facilitated by the whole-body controller, so that operators can focus on the actual task they want to perform. Directed by Prof. Oussama Khatib, a team of 19 researchers from Stanford, the King Abdullah University of Science and Technology (KAUST), and MEKA Robotics are busy building the first prototype, which will be deployed in the Stanford pool in early 2014.  To learn more about Sinbad and the project, please visit theRed Sea Robotics Exploratorium project webpage.

Steering People to Create Vibrant Communities

How do we create and maintain vibrant online communities? How do we motivate people to actively contribute to the community? Prof. Jure Leskovec, together with his PhD student Ashton Anderson and collaborators from Cornell University, has been studying how to steer human behavior in online communities, based on badges. In communities like Wikipedia, Khan Academy, Foursquare, and Stack Overflow, badges are given to users for specific actions or contributions to the community or site. Badges serve as a summary of a user's key accomplishments, and users tend to put in nontrivial amounts of work to achieve particular badges; badges can thus act as powerful incentives. Prof. Leskovec and his collaborators have been studying how best to use badges to reward users for their achievements and contributions. They built a mathematical model that is able to accurately predict how user behavior gets "warped" in the presence of badges. Moreover, their methodology can be used to determine optimal incentives to steer user behavior in a desired direction. Optimal badging leads both to increased participation and to changes in the type of activities that users pursue. Currently, Prof. Leskovec and his collaborators are deploying their methods at the online educational start-up Coursera, with the goal of providing a set of incentive mechanisms that will lead to increased student participation in courses, and lower drop-out rates.

Faculty and Staff Awards and Honors

It's been another banner year for major faculty awards:

• David Dill ― Elected to American Academy of Arts and Sciences
• David Dill again ― Elected to National Academy of Engineering
• Ed Feigenbaum ― IEEE Computer Pioneer Award
• Sachin Katti ― Sloan Foundation Fellow
• Nick McKeown ― ACM SIGCOMM Lifetime Contribution Award
• Subhasish Mitra ― IEEE Fellow
• Eric Roberts ― Karl V. Karlstrom Outstanding Educator Award
• Yoav Shoham ― ACM Fellow
• Yoav Shoham again ― ACM/AAAI Allen Newell Award
• Ryan Williams ― Microsoft Faculty Fellow
• Ryan Williams again ― Sloan Foundation Fellow

In addition, Sebastian Thrun won Smithsonian Magazine's "American Ingenuity Award" in the Education category, Daphne Koller andAndrew Ng were among Time Magazine's "100 Most Influential People," and our beloved long-term department manager Peche Turnerreceived the Staff Leadership Award across the entire School of Engineering. Finally, Mehran Sahami was selected by the Stanford graduating class to be this year’s Class Day speaker, a campus-wide honor never before bestowed upon a CS professor. Mehran delivered his Class Day lecture to a packed Maples Pavilion.

Dissertation Awards

Each year ACM honors the best Computer Science dissertations through its Doctoral Dissertation Awards. Although Stanford didn't capture the big prize in 2012,one of the two Honorable Mentions goes to a Stanford Ph.D., and the other to an incoming faculty member:

• Peter Hawkins' dissertation, Data Representation Synthesis, explores the problem of specifying combinations of data structures with complex sharing and results in provably correct code. In his work on program synthesis, he developed techniques that allow programmers to write their code at a high level, aiding verification while still retaining control over low-level details that are important for efficiency. Peter, who now works at Google, was a student of Prof. Alex Aiken.
• Gregory Valiant’s dissertation, Algorithmic Approaches to Statistical Questions, examines several basic statistical questions from the computational perspective. This work, at the intersection of algorithms, learning, information theory, and statistics, provides insights into several of the challenges encountered in the analysis of today’s large datasets. Greg was advised by Prof. Christos Papadimitriou at UC Berkeley, and he spent a year as a post-doc at Microsoft Research in New England before joining the Stanford faculty in Autumn 2013.

Student News

Our students continue to garner wide recognition for their excellence in research, teaching, and programming.

• Undergraduate Research ― This year's Wegbreit Prize for Undergraduate Computer Science Research went to Bryce Cronkite-Ratcliff for his honors thesis titled “The Development of Automatically Verifiable Systems using Data Representation Synthesis,” advised by Prof. Alex Aiken. Bryce also won a university-wide Firestone Medal for Excellence in Undergraduate Research.
• Graduate Research ― The Arthur Samuel Award for outstanding CS PhD thesis went to Aleksandra Korolova for her thesis titled “Protecting Privacy when Mining and Sharing User Data,”  advised by Prof. Ashish Goel. Aleksandra was advised originally by the late Prof. Rajeev Motwani; we’re grateful to Prof. Goel for jumping in and guiding Aleksandra to such an outstanding thesis. (Prof. Goel is a Stanford CS PhD alum, now a faculty member in Stanford’s Management Science & Engineering department with a “courtesy” appointment in CS.) The Christofer Stephenson Award for the best MS research report went to Tao Wang for “Scene Text Recognition with Convolutional Neural Networks.” His advisors were Profs. Andrew Ng and Daphne Koller.
• Teaching Awards ― The Computer Science Department's Forsythe Teaching Award went to co winners Sebastien Robaszkiewicz and Maria Kazandjieva.  Sebastien, who goes by “Robi,” also won a School of Engineering Centennial TA Award. Robi was the teaching assistant for CS147: Introduction to HCI Design. With 250 students enrolled, CS147 may well have been the largest studio design course taught anywhere. Using a “flipped-classroom” format, Robi was instrumental in designing in-class activities that were engaging and educational across a diverse student population. Maria was the teaching assistant for CS144: Introduction to Computer Networking. Despite being in the throes of finishing her PhD, Maria volunteered to TA CS144 for a third time¾not to earn funding or to satisfy a requirement, but simply due to her love for teaching and helping other students. The students and instructors couldn’t have been happier to have Maria on the CS144 course staff one last time.
• ACM International Collegiate Programming Contest ― Stanford once again advanced to the ACM ICPC World Finals, held in St. Petersburg, Russia this summer.  Our World Finals team, comprised of rising seniors Huaiyu (Nick) Wu, Jack Chen, and Rohan Puttagunta, were invited to the finals after taking first place at the Pacific Northwest Regional Contest. Our team once again made Stanford proud at the World Finals, as they tied for 14th place (with the same ranking as MIT, and solving the same number of problems as some of the bronze medalist teams). CS lecturer  Jerry Cain, who has coached the Stanford team since 2003, was honored by ACM-ICPC with a Coach Award in recognition of bringing Stanford to seven of the last ten World Finals.

Computer Forum News

It’s not only CS majors and enrollments hitting record highs this year. The Computer Forum, our long-standing industrial affiliates program, has reached a new record-breaking membership of 116 companies. Although member benefits include access to both research and recruiting, given the hot hiring climate in the valley, we find many of our newer members are particularly interested in recruiting. This year the Computer Forum helped member companies fill over 300 internship and full-time openings, and hosted hundreds of events, including career fairs, information sessions, company tours, and on-campus one-one interviews.

On the research side, the Annual Meeting was a big success as always, with a theme this year of “Big Data.” In addition to the main meeting, workshops were held in the areas of Security, Mobile & Social Computing, and Information Management. Between the main meeting and the workshops we had over 300 attendees. Many international companies send representatives to the Annual Meeting and workshops each year, but for more extended research interactions, the Computer Forum offers a year-long visiting scholar program. This year we had 20 visiting scholars working directly with faculty and graduate students in a variety of CS research groups.

Now more than ever the Computer Forum is the prime vehicle for companies to recruit students and collaborate in research. If your company is interested in either of these interactions, or better yet both, please consider signing up.

Until the next newsletter, have a terrific year.

Jennifer Widom
Fletcher Jones Professor
Chair, Department of Computer Science